Shape accuracy adjustment of a satellite antenna reflector using RBF Neural Network

In order to improve the shape accuracy of antenna reflector, a non-parametric prediction model based on neural network is proposed in this paper, which is used to control the shape of satellite antenna reflector. The complex system of antenna reflector is regarded as a black box. The RBF neural network model is established by applying voltage to the piezoelectric actuator in advance to adjust the reflector and measuring the change of the reflector surface. The optimal Latin hypercube sampling method is used to obtain the data, and the neural network model is trained. Finally, the reflector is adjusted by particle swarm optimization algorithm. The simulation results show that the on orbit profile control technology based on neural network can significantly improve the surface accuracy of antenna reflector, and does not need an accurate FEM simulation model and other nonlinear parameters of the reflector. The error between the two methods is only 4.67% when the same input voltage of piezoelectric ceramics is brought into the finite element virtual machine.